环境准备

1. 本地客户机

1. idea新建maven工程
2. 配置maven依赖

<dependency>
            <groupId>org.apache.hadoop</groupId>
            <artifactId>hadoop-client</artifactId>
            <version>3.1.3</version>
        </dependency>
        <dependency>
            <groupId>junit</groupId>
            <artifactId>junit</artifactId>
            <version>4.12</version>
        </dependency>

3. 配置的虚拟机hostname, mac环境在/etc/host中配置

2. 虚拟机

1. 本人集群部署NameNode在hadoop102上

文件操作基本流程

Configuration configuration = new Configuration();
FileSystem fs = FileSystem.get(new URI("hdfs://hadoop102"), configuration, "pitaya");
fs.close()

1. 新建configuration对象
2. 新建DistributedFileSystem对象（多态，其父类为FileSystem，可查看源码）
3. 关闭文件流

HDFS文件上传

@Test
public void testCopyFromLocalFIle() throws IOException, URISyntaxException, InterruptedException {
  Configuration configuration = new Configuration();
  //统一资源标识符（Uniform Resource Identifier，URI)
  FileSystem fs = FileSystem.get(new URI("hdfs://hadoop102:8020"), configuration, "pitaya");

  fs.copyFromLocalFile(new Path("/Users/eric/blogs/code/uploadFiles/test.txt"),
                       new Path("/"));
  fs.close();
}

说明

1. pitaya 选择与文件拥有者相同的身份，否则报错(官方文档:Permissions and HDFS)

Permission denied: user=ptaya, access=WRITE, inode="/":pitaya:supergroup:drwxr-xr-x

HDFS文件下载

@Test
public void testDownload() throws IOException, URISyntaxException, InterruptedException {
    Configuration configuration = new Configuration();
    FileSystem fs = FileSystem.get(new URI("hdfs://hadoop102"), configuration, "pitaya");
    fs.copyToLocalFile(true,new Path("/test.txt"),new Path("/Users/eric/blogs/code/downloadFiles"));
    fs.close();
}

HDFS文件更名或移动

fs.rename(new Path("/test.txt"),new Path("/success.txt"));

HDFS文件或目录删除

fs.delete(new Path("/success.txt"),true);

说明

1. 使用时发现FileSystem.get()返回的其实是DistributedFileSystem

   public class DistributedFileSystem extends FileSystem implements KeyProviderTokenIssuer {
   }
   
   //FileSystem里定义
   public abstract boolean delete(Path var1, boolean var2) throws IOException;
   

HDFS文件详情查看

RemoteIterator<LocatedFileStatus> locatedFileStatusRemoteIterator = fs.listFiles(new Path("/"), true);
while(locatedFileStatusRemoteIterator.hasNext()){
    LocatedFileStatus fileStatus = locatedFileStatusRemoteIterator.next();
    System.out.println("======="+fileStatus.getPath()+"======");
    System.out.println(fileStatus.getOwner());
    System.out.println(fileStatus.getModificationTime());
    System.out.println(fileStatus.getBlockSize());
    System.out.println(Arrays.toString(fileStatus.getBlockLocations()));

}

HDFS文件夹和文件创建

fs.mkdirs(new Path("newcreate"));
FSDataOutputStream fsDataOutputStream = fs.create(new Path("/newcreate"));

说明

1. 创建文件夹时发现只能在当前用户目录下进行，而创建文件则没有此要求